И. В. Арнольд лексикология современного английского языка Издание

§ 9.4 SIMILARITY AND DIFFERENCE BETWEEN A SET EXPRESSION AND A WORD
There is a pressing need for criteria distinguishing set expressions not only from free phrases but from compound words as well. One of these criteria is the formal integrity of words which had been repeatedly mentioned and may be best illustrated by an example with the word breakfast borrowed from W.L. Graff. His approach combines contextual analysis and diachronic observations. He is interested in gradation from free construction through the formula to compound and then simple word. In showing the borderline between a word and a formular expression, W.L. Graff speaks about the word breakfast derived from the set expression to break fast, where break was a verb with a specific meaning inherent to it only in combination with fast which means ‘keeping from food’. Hence it was possible to say: And knight and squire had broke their fast (W.Scott). The fact that it was a phrase and not a word is clearly indicated by the conjugation treatment of the verb and syntactical treatment of the noun. With an analytical language like English this conjugation test is, unfortunately, not always applicable.
It would also be misleading to be guided in distinguishing between set expressions and compound words by semantic considerations, there being no rigorous criteria for differentiating between one complex notion and a combination of two or more notions. The references of component words are lost within the whole of a set expression, no less than within a compound word. What is, for instance, the difference in this respect between the set expression point of view and the compound viewpoint? And if there is any, what are the formal criteria which can help to estimate it?
174
Alongside with semantic unity many authors mention the unity of syntactic function. This unity of syntactic function is obvious in the predicate of the main clause in the following quotation from J. Wain, which is a simple predicate, though rendered by a set expression: ...the government we had in those days, when we (Great Britain) were the world’s richest country, didn’t give a damn whether the kids grew up with rickets or not ...
This syntactic unity, however, is not specific for all set expressions.
Two types of substitution tests can be useful in showing us the points of similarity and difference between the words and set expressions. In the first procedure a whole set expression is replaced within context by a synonymous word in such a way that the meaning of the utterance remains unchanged, e. g. he was in a brown study → he mas gloomy. In the second type of substitution test only an element of the set expression is replaced, e. g. (as) white as chalk → (as) white as milk → (as) white as snow; or it gives me the blues → it gives him the blues → it gives one the blues. In this second type it is the set expression that is retained, although its composition or referential meaning may change.
When applying the first type of procedure one obtains a criterion for the degree of equivalence between a set expression and a word. One more example will help to make the point clear. The set expression dead beat can be substituted by a single word exhausted. E. g.: Dispatches, sir. Delivered by a corporal of the 33rd. Dead beat with hard riding, sir (Shaw). The last sentence may be changed into Exhausted with hard riding, sir. The lines will keep their meaning and remain grammatically correct. The possibility of this substitution permits us to regard this set expression as a word equivalent.
On the other hand, there are cases when substitution is not possible. The set expression red tape has a one word equivalent in Russian бюрократизм, but in English it can be substituted only by a free phrase. Thus, in the enumeration of political evils in the example below red tape, although syntactically equivalent to derivative nouns used as homogeneous members, can be substituted only by some free phrase, such as rigid formality of official routine. Cf. the following example:
BURGOYNE: And will you wipe out our enemies in London, too? SWINDON: In London! What enemies?
BURGOYNE (forcible): Jobbery and snobbery, incompetence and Red Tape ... (Shaw).
The unity of syntactic function is present in this case also, but the criterion of equivalence to a single word cannot be applied, because substitution by a single word is impossible. Such equivalence is therefore only relative, it is not universally applicable and cannot be accepted as a general criterion for defining these units. The equivalence of words and set expressions should not be taken too literally but treated as a useful abstraction, only in the sense we have stated.
The main point of difference between a word and a set expression is the divisibility of the latter into separately structured elements which is contrasted to the structural integrity of words. Although equivalent to words in being introduced into speech ready-made, a set expression is different from them, because it can be resolved into words, whereas words are resolved
175
into morphemes. In compound words the process of integration is more advanced. The methods and criteria serving to identify compounds and distinguish them from phrases or groups of words, no matter how often used together, have been pointed out in the chapter on compounds.
Morphological divisibility is evident when one of the elements (but not the last one as in a compound word) is subjected to morphological change. This problem has been investigated by N.N. Amosova, A.V. Koonin and others.] N.N. Amosova gives the following examples:
He played second fiddle to her in his father’s heart (Galsworthy). ... She disliked playing second fiddle (Christie). To play second fiddle ‘to occupy a secondary, subordinate position’.
It must be rather fun having a skeleton in the cupboard (Milne). I hate skeletons in the cupboard (Ibid.) A skeleton in the cupboard ‘a family secret’.
A.V. Koonin shows the possibility of morphological changes in adjectives forming part of phraseological units: He’s deader than a doornail; It made the night blacker than pitch; The Cantervilles have blue blood, for instance, the bluest in England.
It goes without saying that the possibility of a morphological change cannot regularly serve as a distinctive feature, because it may take place only in a limited number of set expressions (verbal or nominal).
The question of syntactic ties within a set expression is even more controversial. All the authors agree that set expressions (for the most part) represent one member of the sentence, but opinions differ as to whether this means that there are no syntactical ties within set expressions themselves. Actually the number of words in a sentence is not necessarily equal to the number of its members.
The existence of syntactical relations within a set expression can be proved by the possibility of syntactical transformations (however limited) or inversion of elements and the substitution of the variable member, all this without destroying the set expression as such. By a variable element we mean the element of the set expression which is structurally necessary but free to vary lexically. It is usually indicated in dictionaries by indefinite pronouns, often inserted in round brackets: make (somebody’s) hair stand on end ‘to give the greatest astonishment or fright to another person’; sow (one’s) wild oats ‘to indulge in dissipation while young’. The word in brackets can be freely substituted: make (my, your, her, the reader’s) hair stand on end.
The sequence of constant elements may be broken and some additional words inserted, which, splitting the set expression, do not destroy it, but establish syntactical ties with its regular elements. The examples are chiefly limited to verbal expressions, e.g. The chairman broke the ice → Ice was broken by the chairman; Has burnt his boats and ... → Having burnt his boats he ... Pronominal substitution is illustrated by the following example: “Hold your tongue, Lady L.” “Hold yours, my good fool.” (N. Marsh, quoted by N.N. Amosova)
All these facts are convincing manifestations of syntactical ties within
176
the units in question. Containing the same elements these units can change their morphological form and syntactical structure, they may be called changeable set expressions, as contrasted to stereotyped or unchangeable set expressions, admitting no change either morphological or syntactical. The examples discussed in the previous paragraph mostly belong to this second type, indivisible and unchangeable; they are nearer to a word than their more flexible counterparts. This opposition is definitely correlated with structural properties.
All these examples proving the divisibility and variability of set expressions throw light on the difference between them and words.